The present disclosure relates generally to control systems for a plant. More particularly, the present disclosure relates to a control system configured to estimate or determine a maximum time constant of a plant for use in configuring the control system.
A plant in control theory is the combination of a process and one or more mechanically-controlled outputs. For example, a plant can be an air handling unit (AHU) configured to control temperature within a building space via one or more mechanically-controlled actuators and/or dampers. Other examples of plants include a chiller operation process, a damper adjustment process, a mechanical cooling process, a ventilation process, a refrigeration process, or any other process in which an input variable to a plant is adjusted to affect an output from the plant.
Many control algorithms for a plant require knowledge about the time scale of the plant being controlled (e.g., the amount of time taken by the plant to execute a process or command). Time scale information is often represented in the form of a single value known as a time constant. However, the use of a single time constant assumes that the plant being controlled can be approximated as a first-order system. If the plant is not behaving as a first-order system, the use of a single time constant estimate may be an issue. Particularly, the time constant may be too small, leading to a sampling interval of the algorithm to be too short. Several techniques have been used to estimate a time constant from data from a plant. Such techniques typically do not work when the responses from the plant are higher order and/or non-linear.
Many algorithms, such as extremum-seeking control (ESC) and proportional-integral (PI) control require an estimate of the plant time constant that relates to the slowest part of a response in a non-linear or higher order system. Standard methods of estimating a time constant do not yield a good estimate in a non-linear or higher order system. There is a need for a method that determines a worst-case or longest time constant estimate from a plant in a non-linear or higher order system.